<html>
<head>
<meta http-equiv="content-type" content="text/html; charset=utf-8">
</head>
<body text="#000000" bgcolor="#FFFFFF">
<p>I'm recycling an old trademark that I've used, <a
moz-do-not-send="true" href="http://caspia.org">Caspia</a>, to
describe my projects to involve Washington State prisoners in
open-source projects. After an afternoon of brainstorming, Caspia
is a new acronym "Creating Accomplished Software Professionals In
Abstentia".<br>
</p>
<p>What does this have to do with Thunderbird? I sat in a room a few
weeks ago with 10 guys at Clallam Bay, all who have been in a
full-time, intensive software training program for about a year,
who are really interested in trying to do real-world projects
rather than simply hidden internal projects that are classroom
assignments, or personal projects with no public outlet. I start
in April spending two days per week with these guys. Then there
are another 10 or so guys at WSR in Monroe that started last
month, though the situation there is more complex. The situation
is similar to other groups of students that might be able to work
on Thunderbird or Mozilla projects, with these differences:</p>
<p>1) Student or GSOC projects tend to have a duration of a few
months, while the expected commitment time for this group is much
longer.</p>
<p>2) Communication is extremely difficult. There is no internet
access. Any communication of code or comments is accomplished
through sneakernet options. It is easier to get things like
software artifacts in rather than bring them out. The internal
issues of allowing this to proceed at all are tenuous at both
facilities, though we are further along at Clallam Bay.<br>
</p>
<p>3) Given the men's situation, they are very sensitive to their
ability to accumulate both publicly accessible records of their
work, and personal recommendations of their skill. Similarly, they
want marketable skills.<br>
</p>
<p>4) They have a mentor (me) that is heavily engaged in the
Thunderbird/Mozilla world.<br>
</p>
<p>Because they are for the most part not hobbyists trying to
scratch an itch, but rather people desperate to find a pathway to
success in the future, I feel a very large responsibility to steer
them in the direction of projects that would demonstrate skills
that are likely to be marketable, and provide visibility that
would be easily accessible to possible future employees. Fixing
obscure regressions in legacy Thunderbird code, with contributions
tracked only in hg.mozilla.org and BMO, does not really fit that
very well. For those reasons, I have a strong bias in favor of
projects that 1) involve skills usable outside the narrow range of
the Mozilla platform, and 2) can be tracked on github.<br>
</p>
<p>I've already mentioned one project that we are looking at, which
is the broad category of Contact manager. This is the primary
focus of the group at WSR in Monroe. For the group at Clallam Bay,
I am leaning toward focusing on the XUL->HTML conversion issue.
Again I would look at this more broadly than just the issues in
Thunderbird, perhaps developing a library of Web Components that
emulate XUL functionality, and can be used both to easily migrate
existing XUL to HTML, but also as a separate library for
desktop-focused web applications. This is one of the triad of
platform conversions that Thunderbird needs to do (the others
being C++->JavaScript, and XPCOM->SomethingElse).</p>
<p>I can see that if the technical directions I am looking at turn
out to work with Thunderbird, it will mean some big changes. These
projects will mostly be done using GitHub repos, so we would need
to improve our ability to work with external libraries. (We
already do that with JsMime but poorly). The momentum in the JS
world these days, unfortunately, is with Node and Chrome V8. That
is going to cause a lot of grief as we try to co-exist with
Node/V8 and Gecko. I could also see large parts of our existing
core functionality (such as the IMAP backend) migrated to a
third-party library.</p>
<p>Our progress will be very slow at first as we undergo internal
training, but I think these groups could start having a major
impact on Thunderbird in about a year.</p>
<p>:rkent<br>
</p>
<p><br>
</p>
</body>
</html>